Preparations of Triphenylmethyl Salts and Their Use in the Cationic

triphenylmethyl salts for the cationic polymerization of a- methylstyrene, with a hope to correlate the counteranion structure and the steric course o...
1 downloads 0 Views 621KB Size
Vol. 8, No. 6, November-December 1975

Preparations of Triphenylmethyl Salts 709

Preparations of Triphenylmethyl Salts and Their Use in the Cationic Polymerization of a-Methylstyrene. Counteranion and Polymer Structure Toyoki Kunitake* a n d Shun'ichi Tsugawa Department of Organic Synthesis, Faculty of Engineering, Kyushu University, Fukuoka, 812, Japan. Received December 5, 1974

ABSTRACT: Triphenylmethyl salts which contain metal chlorides or bromides as the anion were prepared systematically, and employed as initiator for the cationic polymerization of a-methylstyrene. The isotactic unit content of the polymer was found to correlate with the anion size in a relatively simple way except for pentacoordinate anions. This was explicable by considering the steric effect of the counteranion and the coulombic interaction of the growing ion pair. Pentacoordinated counteranions gave rise to greater amounts of the isotactic unit than expected from their sizes, and, in the case of PhxSnCl?Br, the polymer steric structure was dependent on the initiator concentration. These results were attributed to the peculiar structural characteristics of these anions.

Experimental Section

points were measured in sealed capillaries. The analytical results are fairly satisfactory. Repeated purification attempts did not improve the discrepancies in the analysis for some triphenylmethyl salts given in Table I. Ph3CBBr4 was too sensitive to moisture to carry out elemental analysis, and its purity was confirmed from the yield of triphenylcarbinol formed upon hydrolysis.6 Ir spectra were measured as Nujol mulls. Except for the case of Ph3CBC14 and PhsCBBr4 (which decomposed very quickly in the air), all the spectra showed essentially the same pattern and possessed a peak at 1360 cm-' which has been assigned to U C - p h of the triphenylmethyl cation.; NMR spectra were obtained in ethylene dichloride. The measurement was difficult in several cases because of low solubility or facile decomposition; otherwise all the spectra coincided with that reported previously.8 The synthesis of triphenylmethyl salts was also attempted from the following metal halides: HfClj, BiC14, BiBr4, TeCL,, SeCl;, TiC14, and IC13. Although yellow orange powders which undoubtedly contain triphenylmethyl salts were formed, the elemental analytical results were not satisfactory and, therefore, they were not used for polymerization. GeC14 and AsC13 did not appear to react with triphenylmethyl chloride. No triphenylmethyl salts were recovered, when MoCL,, WCls, and VC14 were used. An oily product was obtained from ZnClz and triphenylmethyl chloride. 2. Miscellaneous. The polymerization procedure is the same as described previously.' The steric structure of poly(a-methylstyrene) was determined from the relative peak area of the three methyl signals in the NMR spectrum (chlorobenzene, 12OOC).

1. Preparation of Triphenylmethyl Salts. Metal halides, commercially available and of highest grade, were used without further purification. However, FeCl3 was prepared from Fe and Cln and sublimed three times, and Tic14 and SnC14 were distilled two times. The preparation and characterization of triphenylmethyl salts are summarized in Table I. Most of the triphenylmethyl salts were synthesized from triphenylmethyl chloride, mp lll-112°C (lit.3 110.5-112°C), or triphenylmethyl bromide, mp 149-151.1'C (lit.4 153-154'C), and the corresponding metal halides; triphenylmethyl halide (10 mol% excess) was added to a solution of metal halides (SnC14, SnBr4, GaCly, FeC13, BBrl, BC13) with vigorous stirring under nitrogen in appropriate solvents. Yellow to orange precipitates which formed were filtered, washed with the preparation solvent and/or ether, and dried in vacuo. When the solubility of the metal halides in organic solvents is low (ZrC14, TaC14, NbC16, InBrs), CH2C12 was used as preparation solvent. Unreacted metal halides were removed by filtration and the solvent was evaporated. Orange yellow powders obtained by addition of CC14, followed by vigorous stirring, were washed and dried in vacuo. InC14.HzO and TIC13.4H20 were first stirred vigorously in propionic anhydride and acetic anhydride, respectively, and then equimolar amounts of triphenylmethyl chloride were added as CC14 solution. The precipitates formed were washed with cc14 and dried. Triphenylmethyl perchlorate was prepared from triphenylcarbinol and perchloric acid according to the method of Hoffmann et aL5 Generally, the formation of triphenylmethyl salts was confirmed by elemental analysis and ir and NMR spectroscopy. Melting

Results a n d Discussion 1. Preparation of Triphenylmethyl Salts. In Table 11, the central metals contained in the anion of triphenylmethyl salts are placed in the respective position of the periodic table. In the case of metals shown in parentheses, yellow products are obtained in oil or powder forms upon mixing of triphenylmethyl halides and metal halides. Although these products are undoubtedly triphenylmethyl salts, they could not be isolated in pure form. In the case of the metals which are in italic, triphenylmethyl salts have been obtained only as fluorides. In other cases, the triphenylmethyl salts were successfully prepared as bromide and/or chloride. These salts would be naturally obtainable as fluorides. Alkali and alkali-earth metals do not give the corresponding salts because of the highly ionic nature of the M-X bond. Chlorides or bromides of carbon, silicon, germanium, phosphorus, and arsenic do not form triphenylmethyl salts because their Lewis acidities are too weak. However, the phosphorus and arsenic salts have been formed when fluorides are used.g It is immediately clear from this table that triphenylmethyl salts are most readily formed in the case of counteranions containing the IIIA group metals: B, Al, Ga, In, and T1. The metals belonging to the V group similarly yield triphenylmethyl salts readily. Therefore, triphenylmethyl

The present paper is concerned with the use of various triphenylmethyl salts for the cationic polymerization of amethylstyrene, with a hope to correlate the counteranion structure and the steric course of the cationic propagation. Previously we investigated the cationic polymerization of isobutyl vinyl ether' and a-methylstyrene2 by using some triphenylmethyl salts as initiator. The polymerization is initiated by addition of the triphenylmethyl cation in these systems. As a result, the anion of the initiator salt supposedly becomes the counteranion of the growing chain without structural modification. Therefore, straightforward discussions on the counteranion effect are possible. The previous polymerization results suggested that the size and shape of the counteranion were important factors in controlling the steric structure. However, these studies were conducted with only five triphenylmethyl salts, and, therefore, the conclusion was necessarily of limited nature. In the present study, new triphenylmethyl salts are prepared systematically, and they are used for the polymerization of a-methylstyrene in order to investigate the influence of the counteranion structure on the steric structure of poly(a-methylstyrene).

Macromolecules

710 Kunitake, Tsugawa

Table I Preparation of Triphenylmethyl Salts

_~-_____-__________

_ _ I

Ph,C+XX=

Elemental Anal Preparative solvent

Yield, %

Mp dec, C

Ire

O

NMRe Found (calcd) C, %

H, %

BCI, Cyclohexane 162-169 (lit? 187) 58.28 (57.63) 4.17 (3.82) 78 160-164 (1it.b 182) 46.62 (45.37) as Ph,COH BBr, Cyclohexane GaC1, cc1, 95 167-1 69 0 0 49.61 (50.17) 3.40 (3.32) ( C,H,CO),O-CCl, 70 0 45.17 (45.64) 3.23 (3.00) InC1, TlCl, (CH,CO),O-CCl, 93 153-157 0 0 38.22 (38.72) 2.66 (2.54) 84 161-162 (1it.c 1 7 1 ) 0 0 30.13 (29.96) 2.13 (1.99) SnBr, C6H6 0 0 InBr, CH,Cl, 92 164-1 67 33.77 (33.67) 2.33 (2.23) 0 0 38.96 (39.10) 2.88 (2.59) SnC1,Br CCl, 82 155-161 Sn C1Br, cc1, 24 156-158 0 0 31.69 (31.82) 2.20 (2.10) Zr C1, CH,CI,-CCl, 39 175-1 80 0 44.44 (44.59) 3.18 (2.95) NbCl, CH,@I,-CCI, 82 160-1 66 0 0 41.78 (41.57) 3.06 (2.75) TaCI, CH,Cl,-CCl, 53 162-166 0 33.86 (35.80) 2.55 (2.36) 83 166-1 69 0 51.31 (51.75) 3.72 (3.43) FeCI, CCI, c10, Ac,O 30 133-138 (1it.d 150) 0 0 66.79 (66.60) 4.52 (4.38) a K. M. Harmon, A. B. Harmon, and F. E. Cummings, J. A m . Chem. SOC., 83, 3913 (1961). b Reference 6. c F. Fairbrother and B. Wright, J. Chem. SOC.,509 (1945). d Reference 5. e The circle indicates that ir and/or NMR spectra are satisfactory. Table I1 Central Metals Contained in Triphenylmethyl Salts Prepareda I1 ___A

4 5 6 7

I11 B

A

(Zn) (Cd) (Hg)

B AI Ga In T1

IV B

A

V B

VI B

A

A

B

VI1 A

VI11

PC

(Ti) A SC @e) Fe (Y)b Sn Zr Sbc Nbc (Te) (1) Tac (CeP (Hf) (Bi) (Th)b a The parentheses indicate that impure salts are isolated. The italics indicate that pure triphenylmethyl salts are isolated when metal fluorides (but not chloride and bromide) are employed. b F. Fairbrother, J. Chem. Soc., 503 (1945). c Preparation of triphenylmethyl metal fluorides was reported for these metals: ref 9. salts which contain these metals were mainly employed for the subsequent polymerization study. It is to be noted that metal halides consisting of the metals shown in Table I1 are effective, as such, for the cationic polymerization of vinyl monomers.1° It is also worthy of mentioning that the halogen exchange did not seem t o occur during the preparations of PhsCSnClBr4 and PhsCSnC11Br. On the other hand, Ph3CA1Cl4 was recovered by the reaction of Ph3CCl and AlEtC12. 2. Polymerization. The polymerizations of a-methylstyrene with triphenylmethyl initiators were conducted, as in the previous study, in the following media: (I) 7:3 (v/v) CH2C12-CH3CN; (11) CH2ClZ; (111) 7:3 (v/v) methylcyclohexane-CHzClZ; (IV) 8:2 (v/v) methylcyclohexane-CHZClz; (V) 10:2 (v/v) methylcyclohexane-CH2Cl2. In general, the rate of polymerization increases with the increasing medium polarity, The triphenylmethyl salts were only partially soluble in MCH-CH2C12 mixtures, and, therefore, the initiator concentrations in this medium are apparent values. PhsCFeC14 and Ph3CGaC14 were efficient initiators compared with other triphenylmethyl salts. This is probably attributable to their greater solubilities in the polymerization medium. Table I11 is a summary of the polymerization results in 8:2 methylcyclohexane-CH2ClZ. The steric structure of the polymer was determined from the methyl peak area in the NMR spectrum. Triphenylmethyl salt may dissociate into triphenylmethyl halide and metal halide under the polymerization condition, and there arises a possibility that metal halide initiates polymerization by combination with water and other contaminants, However, this possibility was denied from kinetic evidence in the Ph:$+SnC15--styrene-CH2Ch SYS-

Table I11 Polymerization in 8 : 2 Methylcyclohexane-CH,C1, (Monomer, 0.71 M ; Polymerization Temperature, -78" C; Solvent, CH,Cl, (10 ml) and MCH (40 ml)) Initiator No.

Counteranion

Gon-

- Polymeri- ver-

103M

zation time. hr

sion, %

Ua

44.5 24 1 AICI, 6.05 0.127 161 63b FeCI, 3.10 2 0.129 161 46b 3 GaCI, 4.02 0.1 30 4 SnBr 7.04 0.179 785 0.7 c10, 4.52 5 0.132 785 0.6 106 10 Zr C1 1.33 0.281 6 106 4.2 7 InC1, 8.44 0.127 8 71 3.1 TlCI, 2.76 0.140 9 TaC1, 6.60 0.141 218 9.1 218 10 NbCI, 11.4 0.186 10 218 63b 0.215 11 SnCI, 13.1 12 SnBrC1, 10.9 163 6.3 0.208 378 32b SnClBr, 7.11 0.195 13 0.203 (0.198)C 14 BCI, 14.8 11 183 354 8.7 0.128 (0.138)d InBr, 12.2 15 11 354 0.085 (0.098)e BF, 22.6 16 a The data in parentheses are obtained for the polymer prepared separately under similar conditions. t~ Polymer precipitated noticeably during polymerization. C Initiator concentration: 14.0 mM. d Initiator concentration: 5.42 mM. e Initiator concentration: 31 mM. tem." Sambi e t a1.12 and Bawn e t al.'3 reached the same conclusion. More direct evidence for initiation by the Ph& addition was obtained in the pplymerization of isobutyl vinyl ether.' In this study, we determined the content of the triphenyl-

Preparations of Triphenylmethyl Salts 711

Vol. 8, No. 6 , November-December 1975 0.3

0,3 0,2 3 Y

w 4

0

0.2

0,L

3 Y

0

5

GaCl 4

Y)

3,O 0.1

5-5

4,O

COUNTERANION RADIUS

4,5

(i)

Figure 2. Dependence of the value on the counteranion radius. The initiator concentration is given in parentheses for the pentacoordinate anion.

Figure 1. Variation of the u value with the solvent polarity. The polymerization medium was placed in the order of decreasing polarity ( l / c ) in the abscissa: A, 3:7 CH,CN-CHzClz; B, CHzC12; C, 7:3 MCH-CH2C12; D, 8:3 MCH-CHzClz; E, 8:2 MCH-CHZClz; F, 10:2 MCH-CHZClz.

methyl group by uv spectroscopy, and concluded from a comparison with the number-average molecular weight that 20-100% of the polymer chains are bonded to the triphenylmethyl group. Therefore, the polymerization must be initiated mostly by addition of the triphenylmethyl group if we take into account the facile chain transfer in the cationic polymerization. These results strongly imply that the polymerization was initiated by the addition of the triphenylmethyl cation in the present system as well. The polymer was obtained only when the polymerization medium showed an orange yellow color of the triphenylmethyl cation. It should be noted here that the conversions are greatly different, reflecting the change in the polymerization condition. Therefore, the correlation as described below would be meaningless, if the polymer structure is affected by the conversion. This problem was previously studied by Ohsumi et al.,14 who found that the steric structure of poly(amethylstyrene) was invariant over a wide range of conversion with BF30Et2 in CH2C12. This conclusion would probably apply to the present catalytic system. In fact, the present data, though limited, clearly indicate that the steric structure is constant under a given polymerization condition. 3. Steric Structure of Polymer. The probability of the isotactic propagation u as defined by Bovey was calculated from the fraction of the syndiotactic triad of the methyl signal P,, .2

P,, = (1- u)2 (1) The Bovey plot (triad fraction vs. a) closely fitted the theoretical curve of one parameter u for all the polymer samples, as in the previous cationic polymerizations of this mon ~ m e r . ~ This J~J~ means that the steric course of propagation can be described by the Bernoulli trial.17 The variation of u with some initiators and polymerization solvents is shown in Figure 1. I t is clearly seen that the

u value increases with the decreasing polarity of the polymerization medium. This tendency is the same with initiators having other anions: C104-, NbClG-, and Tack-. The counteranion effect on u becomes more apparent in less polar media. A similar trend was observed previously for different initiaton2 Thus, the influence of triphenylmethyl initiators on the steric structure was extensively studied in the 8:2 MCH-CH2C12 medium. Although the difference in the u value becomes even greater in 10:2 MCH-CH2C12 media, the polymerization was too slow and it was difficult to obtain necessary amounts of the polymer in this solvent. 4. Counteranion Effect on Steric Structure. The u values for the polymers obtained in the 8:2 MCH-CH2C12 solvents are given in Table 111. They fall in the range of 0.1 to 0.2, except for a high value (0.281) for ZrCls-. The difference in the u value between duplicate polymer samples appears to be 0.01 or less. The u values obtained were plotted against the radius of counteranions (Figure 2). The anion radius was estimated, where necessary, from the length of the metal-halogen bond and the van der Waals radius of the halogen atom involved, as mentioned previously.' I t is possible that the effective size of counterions is greater due to solvation. However, this is not likely since Szwarc showed that the conductivity data of (C6H5)3C+SbC15- in CH2C12 were explained without considering specific solvation.ls The counteranions employed appear to be too large for specific solvation. Figure 2 suggests that there exists a fairly simple correlation between the counteranion size and the u value. Excluding the plots of ZrCl5, SnC16, SnBrs, and NbC16, the u value increases with increasing anion size from BF4- to C104- and to BC14-, and then decreases to a fairly constant value for counteranions of greater size. The peculiarity of the pentacoordinated anion was noted before, and is further discussed below. Observation of a greater u value for NbC16- than expected from its size is difficult to rationalize a t the moment. In the cationic polymerization of spiro[2.4]hepta-4,6-diene, NbCl6- showed a counterion effect which is comparable to those of GaC14- and InC14-.19 Further experiments may be required to confirm the validity of these data. In the previous paper we showed for a limited number of the counteranions that the isotacticity of poly(a-methyl-

712

Kunitake, Tsugawa

Macromolecules

styrene) increased with increasing size of the counteranion. It was suggested from these results that the steric effect of the counteranion possessed a predominant influence on the steric course of propagation. A fairly simple correlation observed in Figure 2 again indicates that the anion size is a major factor to determine the steric course for the majority of counteranions. We have proposed the following scheme for the steric course of the cationic propagation.20

Table IV Initiator Concentration and Polymer Structure (Monomer, 0.71 M ; Polymerization Temperature, -78” C; Solvent, CH,CI, (10ml) and MCH (40 ml)) Poly- ConmeriverInitiator zation sion, No. Counterion 103M time, hr % SnBrC1,4.83 183 4.2 0.208 SnBrC1,10.9 163 6.3 0.188 SnBrC1,13.0 378 82.0 0.137,a 0.143b a Fraction which remained in solution during polymerization. b Fraction which precipitated during polymerization. 1 2 3

front-side attack syndiotactic propagation

a’”

I1

S

.

I

-

C

b a c k i d e attack

isotactic

I L

=

conformation reorganization

larger substituent

S = smaller substituent I11 In this scheme, the most stable conformation of the last two units is assumed to be such as shown by I, and the front-side (less-hindered site) attack to the carbonium ion gives rise to a syndiotactic placement and the backside attack an isotactic placement. The relative ease of these modes of monomer attack is decided by the tightness of the growing ion pair for a given monomer; isotactic propagation is preferred with a tight ion pair. The coulombic attraction of the oppositely charged ions descreases with increasing ion sizes, whereas the steric hindrance of the counteranion toward monomer attack may simply increase with its size. The observed tendency is explicable by a combination of these two effects. The front-side attack becomes relatively less favorable as the anion size increases from BF4- to BC14- because of their increasing steric hindrance. Further increase in the anion size may decrease the tightness of the ion pair due to decreased electrostatic attraction, provided that the steric influence levels off. Independent support of this interpretation may be obtained by the investigation of the steric course of @-substitutedmonomers which is being studied at present in these laboratories. The conformation of the growing chain end will change with different counterions and solvents. However, the conformation change cannot affect the relative arrangement of the penultimate and terminal substituents and the counterion in the above scheme.20 In the cationic polymerization of isobutyl vinyl ether, considerable difference in the steric structure was caused by the use of different IIIA group metal chlorides as counteranion.’ The corresponding change in the avalue of poly(cymethylstyrene) is barely beyond the experimental error except for BC14-, although the trends in these two systems are the same.

The character of counteranions may also be influenced by the nature of the metal-halogen bond involved. Therefore, the ionization potential or electronegativity of the metal involved may be utilized as indices of the counteranion character. Unfortunately, the variation of the u value was relatively small, if we exclude the pentacoordinate anions. Therefore, these correlations will be discussed in other polymerization systems where the difference in the steric structure is more pronounced. 5. Pentacoordinated Counteranions. It was indicated in the previous papers that the use of Ph3CSnC15 as initiator gave rise to much greater amounts of the isotactic unit than expected from its counterion size.’,* Analogous results were obtained for SnClBr4, SnCldBr, SnBr5, and ZrC15, as seen from Figure 2. As mentioned before, these data are most satisfactorily explained by considering the peculiar structure of these counteranions. Pentacoordinate compounds generally assume non-rigid structures with facile interconversion between trigonal bipyramid and tetragonal pyramid structures.21 The conceivably asymmetrical nature of these pentacoordinate counteranions may lead to tighter ion pairs than expected otherwise. An interesting observation related to this discussion is the dependency of the steric structure on the initiator concentration shown in Table IV. The u value decreased from 0.208 to 0.14 with increasing initiator concentrations. This u variation cannot be attributed to precipitation of the polymer during polymerization, since the two fractions in run No. 3 gave the identical u values within the experimental error. Pentacoordinate anions are capable of association by using the vacant coordination site, unlike tetra- and hexacoordinated compounds. Therefore, it is possible that the structure of pentacoordinate counteranions varies by changing their concentrations. The results given in Table IV seem to support this view. The decreasing u value with increasing PhSCSnCldBr may indicate that the nature of this pentacoordinate anion approaches those of fully coordinated counteranions. In fact, the u value obtained with the highest PhsCSnC14Br concentration is 0.14 and this value is in the same range as those obtained with fully coordinated counteranions of comparable size. The change in the initiator concentration did not appear to affect the u value in the case of BF4, BC14, and InBr4 counteranions (see Table 111), in accordance with the above theory. The present polymerization system is not advantageous for the elucidation of this problem, s i w e the variation of the steric structure is not necessarily great. More definite discussions may be done in the cationic polymerization of vinyl ethers in later publications.

Acknowledgment. We are grateful to Mr. K. Takarabe for carrying out a part of the experimental work.

Vol. 8, No. 6, November-December 1975

Mechanism of Radical Copolymerizations of N-Vinylcarbazole 713

References and Notes (1) T. Kunitake, Y. Matsuguma, and C. Aso, Polym. J . , 2,345 (1971). ( 2 ) Y. Matsuguma and T. Kunitake, Polym. J.,2,353 (1971). (3) W. E. Bachmann, “Organic Syntheses”, Collect. Vol 3, Wiley, New York, N.Y., 1955, p 841. (4) K. D. Berlin, L. H. Gower, J. W. White, D. E. Gibhs, and G. P. S t u r m , J. Org Chem., 27,3595 (1962). ( 5 ) K. A. Hoffmann and H. Kirmreuter, Ber., 42,4861 (1909). (6) K. M. Harmon and A. B. Harmon, J . Am. Chem. Soc., 83,869 (1961). ( 7 ) R. E. Weston, J r . , A. Tsukamoto, and N. N. Lichtin, Spectrochim. Acta, 22,433 (1966). (8) D. G. Farnum, J . Am. Chem. Soc., 86,934 (1964). (9) D..W. A. Sharp and N. Shepperd, J. Chem. Sac., 674 (1957). (10) E.g., T. Higashimura, “Cationic Polymerization”, Kagakudojin, Kyoto, 1971, p 23. (11) T. Higashimura, T. Fukushima, and S. Okamura. J . Macromol. Sci.,

Chem., 1,683 (1967). (12) M. S. Sambi and F. E. Treloar, Polym. Lett., 3, 445 (1965); M. S. Sambi, Macromolecules, 3,351 (1970). (13) C. E. H. Bawn, C. Fitzsimmons, A. Ledwith, J. Penfold, D. C. Sherrington, and J. A. Weightman, Polymer, 12, 119 (1971). (14) Y. Ohsumi, T. Higashimura, and S. Okamura, J . Polym. Sci.. Part A - 1 . 4,923 (1966). (15) S . Brownstein, S. Bywater, and D. J. Worsfold, Makromol. Chem.. 48, 127 (1961). (16) Y. Ohsumi, T. Higashimura, and S. Okamura, J . Polym. S c i . . Part A , 3. 3729 (1965). (17) F. A. Bovey, “High Resolution NMR of Macromolecules”, Academic Press, New York N.Y., 1972, p 151. (18) N. Kalfoglou and M. Szwarc, J . Phys. Chem., 72,2233 (1968). (19) T. Kunitake, T. Ochiai, and 0. Ohara, J . Polym. Sci., Polym. Chem. Ed.,in press. (20) T. Kunitake and C. Aso, J . Polym. Sci., Part A - I , 8,665 (1970). (21) E. L. Mutterties and R. A. Schunn, Q . Reu., Chem. SOC.,245 (1966).

Mechanism of Radical Copolymerizations of N-Vinylcarbazole with Diethyl and Dimethyl Maleate Motomu Yoshimura, Hiroshi Mikawa, * and Yasuhiko Shirota Department of Applied Chemistry, Faculty of Engineering, Osaka University, Yamada-Kami, Suita, Osaka, 565, Japan. Received July 18,1975

ABSTRACT: Free-radical initiated copolymerizations of N-vinylcarbazole (VCZ) (MI) with diethyl maleate (Mp) or dimethyl maleate (M2’) have been studied in connection with the 1:l alternating radical copolymerization of VCZ with the isomeric diethyl fumarate. The compositions of the copolymers vary from 1:l alternation to increasingly VCZ rich with the increase of VCZ in the monomer feed (Table I). Results are strongly in favor of the mechanism involving the participation of free monomers in the alternating copolymerization (eq 2-5). Copolymer composition is well reproduced with rl = 0.27 (Mz), 0.26 (M2’); rp = 0. This mechanism leads to the copolymerization rate eq 14,which fits well the experimental data (Figure 2).

The mechanism for the alternating radical copolymerization has recently attracted a great deal of attention, and is still controversial. Some propose the charge-transfer model to explain the alternation, that is, the charge-transfer complex formed between monomer pairs (the charge-transfer complexed monomer) homopolymerizes to yield the 1:l alternating copolymer.1 Others insist that the cross-propagation of free monomers, the transition state of which is stabilized, is operative.2 We have been investigating free-radical initiated copolymerizations of N-vinylcarbazole (VCZ), a typical electrondonating monomer which has recently received great attention in the field of the so called charge-transfer polymerization, with a series of electron-accepting monomers with a hope of elucidating the mechanism of the alternating radical copolymerization. In the previous paper,3 we reported the first known examples of the 1:l alternating radical copolymerization of VCZ with diethyl fumarate or fumaronitrile, and proposed a new model for the alternating radical copolymerization. The generalized mechanism which we have proposed involves the participation of both free monomers and the charge-transfer complexed monomer (CT-complexed monomer) in the propagation step of the polymerization with a varying degree of their contributions to the polymerization which depends on the reaction system. Our model stands on the basis that the alternation can be brought about by the stabilization of the transition state of the cross-reaction of both free monomers and the CT-complexed monomer, and that the participation of the CTcomplexed monomer in the polymerization makes contri-

bution to the enhancement of the polymerization rate due to its higher reactivity than the free monomer. Our continued interest in this field has led us to extend the study to the copolymerization of VCZ with the isomeric electron-accepting monomer of diethyl fumarate and diethyl and dimethyl maleate. The present results are consistent with our model and seem to provide the experimental evidence for the participation of the free monomer in the alternating radical copolymerization.

Experimental Section Materials. N - Vinylcarbazole (VCZ) and 2,2’-azobisisobutyronitrile (AIBN) were purified and dried as described in the previous paper.4 Diethyl maleate and dimethyl maleate which were obtained commercially (WAKO Pure Chemical Ind. Ltd.) as pure grade reagents were dried over calcium chloride, and distilled twice from calcium hydride in vacuo under a nitrogen stream, bp 71.5OC ( 2 mm) and 60.5OC (2 mm), respectively. The purity of the material was checked by means of GLC. Polymerization Procedure. All copolymerizations were carried out in a benzene solution using AIBN as an initiator. The total monomer concentration was fixed as 1.0 M, the monomer feed ratio being varied widely. The initiator concentration was 1.0 X IO-:{ M . The polymerization solution contained in a glass tube, ca. 1.5 cm in diameter, was evacuated a t 2 X Torr by means of several freeze-pump-thaw cycles, sealed off and then allowed to stand for an appropriate time in a thermostat maintained a t 80 f 0.1OC. The polymerization was stopped a t low conversions. The polymerization solution was chilled and poured into a large volume of MeOH to precipitate the polymer. The polymers were collected by filtration with a glass filter, washed with MeOH repeatedly, dried in vacuo, and weighed. The polymers were reprecipitated twice from benzene with MeOH and submitted to analyses.